Loading...

Frequently Asked Questions for the Fujitsu ETERNUS Storage Array

What are the precedence rules for selecting the target volume?

For Fujitsu ETERNUS Clones

Precedence is given to any aged clone on the same source logical unit number (LUN) that is available on the storage array.

If you select the snap configuration property  Enable Incremental Clone, the Commvault software does not physically delete the clone from the storage array during the delete operation. Instead, the Commvault software renames the target volume to indicate logical deletion. The deleted clone name has the following format:

iSnapR_#<VolumeNo>or <Label>_#<VolumeNo>, where Label is the snap configuration property Volume Label Prefix.

Any subsequent request for a clone backup operation on the same source LUN runs an incremental sync backup operation on the deleted (aged) clone.

If no aged clone is found on the same source LUN, and if you specified a prefix in the snap configuration property Volume Label Prefix, then the preference is given to target volumes whose name has a matching prefix. If there is more than one matching volume, the source volume type defines the precedence:

  • For standard source volumes, standard target volumes are selected.
  • For all other source volumes, thin provisioning volumes are selected.
  • If there is no matching volume, the backup job will fail.

If you didn't specify a prefix in the Volume Label Prefix property, then a new target volume is created in the destination pool. If no destination pool is configured, a new target volume is created in the source pool.

For Fujitsu ETERNUS Snap and Fujitsu ETERNUS SnapOPC+ Snapshots

If you specify a volume label in the Volume Label Prefix snap configuration property, SnapOPC and SnapOPC+ snapshots use any snap data volume (SDV) whose name has a matching label prefix.

If there is no matching SDV available with a name that contains the prefix, then the backup job will fail.

How does the Commvault software delete SnapOPC+ snaps?

The Fujitsu ETERNUS storage array uses a chain of SnapOPC+ snapshots that is sorted by timestamp in ascending order (the oldest first) to manage the sequence of snapshots.

Due to the design of the Fujitsu ETERNUS storage array, if an SnapOPC+ snapshot is deleted out of the sequence on the chain, then all older snapshots are deleted, even if the snapshots are still active .

To prevent the deletion of active snapshots on the storage array, the Commvault software deletes SnapOPC+ snapshots in two steps:

  1. The snapshot is logically marked as deleted and the Commvault software renames the target SDV (Snap Data Volume) to indicate logical deletion using the following name format:

    iSnapD_#<VolumeNo>

  2. In the second step, the Commvault software scans the entire SnapOPC+ chain of snapshots on the source LUN. Based on the volume label and the timestamp, the Commvault software verifies if the snapshot could cause an out-of-sequence deletion on the storage array. If the source LUN has at least one older SnapOPC+ snapshot not yet marked for deletion, then the delete operation deletes the current snapshot from the CommServe database, but not from the storage array.

A SnapOPC+ snapshot can be physically deleted from the storage array when the source LUN does not have any SnapOPC+ snapshots that are older than the snapshot being deleted or if one of the following conditions is true:

  • The snapshot being deleted is the first snap of  the chain in the source LUN.
  • All older SnapOPC+ snapshots in the chain are marked for deletion by  the Commvault software.

The following are a few examples of the SnapOPC+ snapshot deletion. 

Initial SnapOPC+ chain of snapshots

Request

Resulting SnapOPC+ chain of snapshots

S1, S2

A delete request for snapshot S1 physically deletes S1 from the storage array because there is no older snapshot than S1 on the storage array.

S2

S1, S2

A delete request for S2 will delete S2 from CommServe database, but not from the storage array because there is an older snap S1 still active.

The Commvault software marks the S2 logically deleted on the array.

S1, S2(LD)

where LD stands for logically deleted

S1, S2(LD)

A delete request for snapshot S1 causesboth S1 and S2 to be physically deleted on the storage array, since S2 is already marked for deletion.

 

S1, S2, S3, S4, S5

An out-of-sequence delete request for snapshots S2, S3, and S5 .

S1, S2(LD), S3(LD), S4, S5(LD)

S1, S2(LD), S3(LD), S4, S5(LD)

A delete request for snapshot S1 deletes snapshots S1, S2, and S3, since S2 and S3 are already marked for deletion.

S4, S5(LD)

The target volume of all the physically deleted sessions will be renamed as the following:

  • If you specified the snap configuration property Volume Label Prefix, then the target volume name is VolumeLabelPrefix_#<LUN No>. If you didn't specify the snap configuration property Volume Label Prefix, the target volume has the following format: iSnapD_#<VolumeNo>.
  • The target volume of logically deleted SnapOPC+ snaps is iSnapD_#<VolumeNo>.

See the ETERNUS documentation on the storage vendor website for more details on SnapOPC+ snapshots.

Last modified: 3/14/2018 9:50:17 PM